Remote ischemic conditioning counteracts the intestinal damage of necrotizing enterocolitis by improving intestinal microcirculation

Necrotizing enterocolitis (NEC) is a devastating disease of premature infants with high mortality rate, indicating the need for precision treatment. NEC is characterized by intestinal inflammation and ischemia, as well derangements in intestinal microcirculation. Remote ischemic conditioning (RIC) has emerged as a promising tool in protecting distant organs against ischemia-induced damage. However, the effectiveness of RIC against NEC is unknown. To address this gap, we aimed to determine the efficacy and mechanism of action of RIC in experimental NEC. NEC was induced in mouse pups between postnatal day (P) 5 and 9. RIC was applied through intermittent occlusion of hind limb blood flow. RIC, when administered in the early stages of disease progression, decreases intestinal injury and prolongs survival. The mechanism of action of RIC involves increasing intestinal perfusion through vasodilation mediated by nitric oxide and hydrogen sulfide. RIC is a viable and non-invasive treatment strategy for NEC.

A1 vessels. Anatomic assessment of vessel size does not necessarily translate into changes in nutritive perfusion and is not a quantitative measure of flow. Similarly, changes in feed artery flow could be directed to tissue perfusion or be diverted through AV shunts, without improving perfusion. Use of microspheres or even laser Doppler assessment of flow would better address these issues. 2. Arteriole diameter and velocity and flow volume are measured at baseline. To adequately assess whether perfusion is sufficient, some stress is needed (e.g. baseline and max flow following adenosine infusion). Loss of flow reserve would better indicate an ischemic environment. Alternatively, measuring venous oxygen saturation or lactate production would help establish presence of ischemia. 3. The bulk of the presented data are observational. Only figure 4 begins to address mechanism but the data are not as robust as expected. H2S synthase inhibitors make NEC grade worse but no data are provided to suggest this is related to changes in flow. Administration of H2S reduces NEC. However there are key missing experiments. First, can H2S rescue the effect of H2S synthesis inhibitors? Second, what is happening to H2S levels in the tissue during these interventions? Third, does scavenging of H2S (e.g. vitamin B12) recapitulate the effects of H2S synthase inhibitors? The authors could also use genetic models devoid of H2S inhibitors to more selectively test their role in NEC. Finally there is no control for the potent vasodilating effects of H2S. Use of NO, adenosine, or nitroprusside in figure 4e would help establish specificity for H2S. 4. It is presumed by the authors that the mechanism of early and late RPC is the same since they are equally effective. This logic is flawed and only assessment of early RPC mechanism can be claimed.
5. An important missing control is to determine the effect of RIC and H2S inhibitors in control breast fed animals.
Other comments -Much of the results under "mechanism" are really just further characterization of the NEC process. For example, changes in vascular density, height, leukocyte adhesion, arteriolar diameter and flow are all phenotypical changes that occur with NEC and may be improved with IPC but no data are presented to show they are mechanistic rather than simply coincident with NEC.
Reviewer #3: Remarks to the Author: This report studies RIC in a murine model of neonatal enterocolitis. RIC is tested "early" (pre RIC days 5 and 7) and "late" (Days 6 and 8; per and post RIC) RIC is remarkably effective and the mechanism involves improved microperfusion. Improved perfusion with RIC has been shown previously in both brain and coronary circulations as the authors reference. The methods are elegant. This work is highly translatable as the authors add to the evidence that RIC is safe and effective Some questions 1.) Was RIC applied on one or both hindlimbs? Not clear in lines 305-8 2.) How was this regimen of RIC chosen? They were certainly effective but how were the regimens arrived at? 3.) The authors studied the H2S system. The NO and NOS system has also been implicated in the mechanism of action of RIC and are important in perfusion. Were NOS 3 inhibitors tested or NOS3 KO mice? The authors might comment on this 1. ischemic preconditioning has been shown to reduce inflammatory responses in many animal models of sepsis and inflammation (see for ex PMID: 30809283, PMID 28437377, PMID 26436208, PMID 24904237, PMID 25037959) among others; these papers offer significant mechanistic insights that could be tested in the current model.
We appreciate this comment from the reviewer. In the present study, we have identified that the mechanism of action of RIC is the restoration of intestinal perfusion through enhanced vasodilation. We have validated this by using inhibitors of the vasodilatory gasotransmitter, hydrogen sulfide (H2S), H2S scavengers, and inhibitors of Nitric Oxide (NO), the downstream effector of hydrogen sulfide (Fig. 5, 6). In addition, we investigated the role of NO in the mechanism of action of RIC by studying the effect of vasodilation in eNOS knockout mice ( Supplementary Fig. 5). Our findings demonstrate a critical role for the vasodilatory action of H2S and NO in the RIC-mediated preservation of intestinal perfusion, leading to a reduction in intestinal injury and inflammation and enhancing survival during experimental NEC (lines 199-268).
2. The finding of improved perfusion is consistent with the finding that NEC is improved. Missing is evidence that the improved perfusion is the reason for the improvement in NEC, and not the result of it.
Thank you for this comment. We found that the mechanism of action of RIC is restoration of intestinal perfusion during experimental NEC, which is mediated through 2 vasodilation and leads to an improved outcome of NEC. In order to validate this, we have targeted the vasodilatory action of hydrogen sulfide (H2S), which has been demonstrated previously to have a critical role in improving intestinal perfusion during experimental NEC 1 . We have used inhibitors of endogenous H2S-synthesizing enzymes, H2S scavengers, as well as inhibitors of NO. Our findings demonstrate that inhibition of vasodilation renders RIC ineffective in improving intestinal wall perfusion ( Fig. 5a-b) and intestinal microcirculation (Fig. 5d-e, g-h). As a result of this, NEC-induced intestinal injury and inflammation are also not improved (Fig. 6), ultimately leading to the same mortality observed in NEC alone (Fig. 7). These findings provide evidence that the RIC-mediated improvement in intestinal perfusion is not secondary to improved intestinal morphology and is required to counteract the effects of NEC (lines 242-268). 3. A large, multicenter trial published in the nejm in which 1403 patients were studied for the role of remote preconditioning for heart surgery showed zero benefit for patients. this raises doubts regarding the rationale, and the significance, of the cur6ent study.
The reviewer is right in pointing out that the effect of preconditioning in improving outcome after heart surgery is controversial.
The rationale of our study is based on the fact that necrotizing enterocolitis (NEC) has an intestinal ischemic component in its pathophysiology. RIC has been applied in many different settings in both humans and animals, in which ischemic injury was involved. Clinical trials have been performed in adults 2-4 and in children [5][6][7][8][9] indicating benefits from RIC in various organs including heart, lung, and kidney. In addition, a systematic review and meta-analysis evaluating randomized trials, found that compared with controls, RIC significantly reduced the recurrence of stroke or transient ischemic attacks 10 . However, the advantage of RIC remains controversial as two large trials have shown no improvement in relation to cardiac surgery 11,12 . However, only two trials in adults have focused on the effects of RIC on the intestine. One trial indicated benefit after abdominal aortic aneurism repair when the intestinal ischemia/reperfusion injury is expected 13 and the other trial indicated no intestinal changes after cardiopulmonary bypass when the intestinal injury is rare, moderate and transient 14 . To our knowledge, the potential benefits of RIC in preterm infants and particularly in those with NEC have not been investigated. Our results indicate that RIC is remarkably effective in blunting intestinal ischemic damage in neonatal pups, justifying further investigation of its effectiveness in human preterm infants.
To further clarify our rationale, we added a section in the Introduction, summarizing the information above (lines 96-106).
4. Various mechanisms are proposed to explain the benefit of RIC on page 5at least one of these could be tested in the current model, strengthening the current work.
Thank you for this comment. Previous experimental studies in the heart and the brain have demonstrated that remote ischemic conditioning targets the microcirculation in distant ischemic organs [15][16][17][18] . According to these findings, we have demonstrated that RIC improves intestinal perfusion and microcirculation via the action of endogenous vasodilatory gasotransmitters including hydrogen sulfide and nitric oxide. To further explore the mechanism of action of RIC ( Fig. 5-7), we performed additional experiments using inhibitors of hydrogen sulfide and nitric oxide as well as eNOS knockout mice. We demonstrated that inhibition of vasodilation eliminates the protective effect of RIC characterized by normalization of the intestinal perfusion (Fig. 5), elimination of the intestinal epithelial damage and inflammation (Fig. 6), and ultimately, improvement in survival ( Fig. 7) (lines 199-268). Fig. 1 in the study design, the early RIC actually occurs after the first time point in the late ric group. this makes interpretation difficult as to what is early vs. late.

5.
We appreciate this comment from the reviewer. Previous studies have suggested that remote ischemic conditioning activates two distinct time frames of protection against ischemia reperfusion (IR) injury in the brain and heart. The initial window of protection occurs immediately after the RIC stimulus and lasts for 2 hours, whereas the second window of protection occurs 12-24 hours after the RIC stimulus and lasts 48-72 hours 19,20 . Moreover, it has been reported that compared with a single episode of remote ischemic conditioning, repeated episodes were more protective in reducing inflammation in the ischemic myocardium 21 . Based on such findings, we chose distinct time points of RIC, 48 hours apart, through the course of our disease model. 7. Fig. 1i the survival data is potentially profound, but it looks as though the differences are reflective of a small number of pupsplease address how many pups.
We appreciate this critical comment from the reviewer. In our initial experimental design, pups were sacrificed at P9. We increased the sample size to include at least 20 pups per group and demonstrated that survival up to P9 was enhanced by either αRIC or βRIC (Fig. 7) in wild type. However, this beneficial effect was not seen in wild type pups given inhibitors of H2S (Fig. 7), or in eNOS knockout pups ( Supplementary Fig. 5d).
In addition, to further characterize the outcome after RIC, we performed additional experiments aimed at quantifying the survival rate after P9. Pups in the various experimental groups continued to receive gavage feeding and were observed by an investigator blinded to treatment allocation until death occurred. Compared to NEC alone, the survival was extended by either αRIC or βRIC (Fig. 7). Conversely, inhibitors of H2S resulted in a similar mortality rate to NEC alone. These important findings reinforce our observations ( Fig. 5 and 6) on the mechanism of action of RIC being dependent on the H2S pathway. (Fig. 7) (lines 256-266).
8. Fig. 2 the middle panel appears to be cut in a very different plane than all other panels; the perfusion index doesn't seem reflective.
We have corrected the middle panel to match the others and added indicators in Fig. 2. In addition, we clarified that the perfusion index refers to the ratio of the intra-villi arteriole to the whole villi (lines 494-499) indicating significant differences between NEC alone and NEC with RIC.
9. Fig. 3 these images are beautiful but do not add to the underlying mechanism. the mech by which RIC improves perfusion is not tested.
In the revised manuscript, we have provided further evidence to explain the underlying mechanism of action of RIC by performing various new experiments. Our findings suggest that RIC improves the outcome of NEC, enhances survival, and restores the NEC-induced derangements in intestinal perfusion via a vasodilatory-dependent mechanism. We have demonstrated this by assessing changes in the RIC-mediated restoration of intestinal perfusion upon treatment with inhibitors of vasodilatory mediators such as hydrogen sulfide and nitric oxide. Our findings suggest that inhibition of endogenous synthesis of hydrogen sulfide through administration of H2S-synthesizing enzyme inhibitors, and scavenging of H2S, both considerably abolish the protection conveyed by RIC during experimental NEC, leading to impaired perfusion, increased intestinal injury and inflammation, and poor survival ( Fig. 5-7). Our findings also suggest that inhibition of NO-mediated vasodilation abolishes the RIC-mediated improvements in intestinal wall perfusion and increases the intestinal injury. Taken together, the mechanism by which RIC improves intestinal perfusion during experimental NEC is the collective vasodilatory action of gasotransmitters such as H2S and NO.
We have added the above explanation in the manuscript (lines 199-268).

Reviewer #2 (Remarks to the Author):
This study examined the effect of RPC on vascular function in an induced murine model of NEC. Conceptually there is modest novelty since RIC has previously been applied to protection of gastric ischemia and improved mucosal blood flow. Here the role of H2S is examined. Several major concerns limit enthusiasm for this paper.
Major concerns 1. Vascular function and ultimately tissue perfusion are the key outcome measures in this study. However this is measured by anatomic assessment of vessel density and volumetric flow through A1 vessels. Anatomic assessment of vessel size does not necessarily translate into changes in nutritive perfusion and is not a quantitative measure of flow. Similarly, changes in feed artery flow could be directed to tissue perfusion or be diverted through AV shunts, without improving perfusion. Use of microspheres or even laser Doppler assessment of flow would better address these issues.
We thank the reviewer for the comments. To assess changes in arterial flow, we measured flow velocity in the entire intestinal wall using Doppler Ultrasound 25,26 .
Intestinal wall perfusion was calculated as average flow velocity (mm/s) of multiple abdominal regions (Fig. 4a). In agreement with intestinal damage (Fig. 1a) and decreased microcirculation (Fig. 3), intestinal wall flow velocity ( Fig. 4) was significantly reduced in NEC pups, compared to breastfed controls. In contrast, both αRIC and βRIC increased flow velocity in the intestinal wall ( Fig. 4) demonstrating improved intestinal perfusion (Supplementary movies 9-12). Conditioning with αRIC or βRIC in breastfed control pups did not alter intestinal wall flow velocity ( Supplementary Fig. 1f). We added these results in the manuscript (lines 187-195).
These findings are in agreement with the changes observed in intestinal microcirculation using two photon laser scanning microscopy (TPLSM) (lines 187-195).
2. Arteriole diameter and velocity and flow volume are measured at baseline. To adequately assess whether perfusion is sufficient, some stress is needed (e.g. baseline and max flow following adenosine infusion). Loss of flow reserve would better indicate an ischemic environment. Alternatively, measuring venous oxygen saturation or lactate production would help establish presence of ischemia. 6 As suggested, we have performed additional experiments using two photon laser scanning microscopy (TPLSM) to assess submucosal arteriole diameter, velocity, and flow volume at baseline in P5 and P9 pups in response to formula feeding as a stress factor ( Fig. 3b-d). Our findings indicate that single formula feeding at P5 caused no significant change in arteriole velocity, diameter, and flow volume. On the contrary, later in the neonatal period (P9), formula feeding resulted in significant increase over baseline in arteriole velocity, diameter, and flow volume. Taken together, these findings suggest that the immature submucosal arterioles of the intestine respond poorly to feeding in early neonatal mice (P5), which could contribute to feeding-induced intestinal hypoxia and development of NEC. Excitingly, RIC is able to counteract the poor response to feeding in early neonatal mice by improving intestinal microcirculation, leading to protection against NEC development and ultimately enhanced survival. Please see We agree on the importance of assessing intestinal ischemia. Pimonidazole is a sensitive marker of intestinal ischemia and most importantly it allows the localization of ischemia 27 . In accordance with previous work from our group 27 , we found that NEC is associated with ischemia at the tip of the villi (Fig. 2). We performed additional experiments and discovered lack of vascular flow at the tip of the villi explaining the ischemia occurring in this area. RIC during experimental NEC improves the flow to the top of the villi, thus avoiding ischemia (Fig. 2).
3. The bulk of the presented data are observational. Only Fig. 4 begins to address mechanism but the data are not as robust as expected. H2S synthase inhibitors make NEC grade worse but no data are provided to suggest this is related to changes in flow. Administration of H2S reduces NEC. However there are key missing experiments. First, can H2S rescue the effect of H2S synthesis inhibitors? Second, what is happening to H2S levels in the tissue during these interventions? Third, does scavenging of H2S (e.g. vitamin B12) recapitulate the effects of H2S synthase inhibitors? The authors could also use genetic models devoid of H2S inhibitors to more selectively test their role in NEC. Finally, there is no control for the potent vasodilating effects of H2S. Use of NO, adenosine, or nitroprusside in Fig. 4e would help establish specificity for H2S.
We appreciate the above comments from the reviewer, and we have performed various additional experiments to address these questions.
(1) We investigated whether administration of NaHS, an exogenous H2S donor, can rescue the effect of H2S synthesis inhibitors. In Fig. 5 and 6, following treatment with H2S synthesis inhibitors, NaHS did not improve intestinal morphology and did not reduce intestinal inflammation. Likewise, NEC-induced derangements in intestinal wall perfusion and impaired velocity, diameter, and flow volume of submucosal arterioles were not rescued (lines 199-268). (2) We have performed additional experiments to validate changes in expression of H2S in the ileum during these interventions. Study of H2S under in vivo conditions is challenging due to the short half-life of H2S 28 . In addition, following our consultation with experts in the field, we learned that direct measurement of H2S levels in tissue is not an ideal approach. Hence, we performed immunofluorescence staining for cystathionine-β-synthase (CBS), one of the key endogenous H2S-synthesizing enzymes in the ileum. Please see Supplementary 4. It is presumed by the authors that the mechanism of early and late RPC is the same since they are equally effective. This logic is flawed and only assessment of early RPC mechanism can be claimed.
We have modified the nomenclature of early and late RIC to αRIC (P5 and P7) when the intestinal damage was not yet present or it was minimal; and βRIC (P6 and P8) when changes in intestinal damage start to be detectable. We performed additional experiments to evaluate the effects of αRIC as well as βRIC. Our results demonstrated that αRIC and βRIC were equally effective in NEC and had the same mechanism of action ( Fig. 5-7).
5. An important missing control is to determine the effect of RIC and H2S inhibitors in control breast fed animals. 8 We have performed additional experiments and studied the effect of RIC and H2S inhibitors in breastfed controls. We have not observed any change in intestinal morphology, inflammation, and intestinal perfusion by RIC on breastfed control pups. These findings are shown in Supplementary Fig 1. Other comments -Much of the results under "mechanism" are really just further characterization of the NEC process. For example, changes in vascular density, height, leukocyte adhesion, arteriolar diameter and flow are all phenotypical changes that occur with NEC and may be improved with IPC but no data are presented to show they are mechanistic rather than simply coincident with NEC.
Thank you very much for your comments. We have found that the mechanism of action of RIC is restoration of intestinal perfusion during experimental NEC, which is mediated through vasodilation and leads to improved NEC outcome. To validate this, we have targeted the vasodilatory action of H2S, which has been demonstrated by previous authors to have a critical role in improving intestinal perfusion during experimental NEC 1 . We have used inhibitors of endogenous H2S synthesizing enzymes, H2S scavengers, as well as inhibitors of nitric oxide, a downstream effector of H2S. Our findings demonstrate that inhibition of vasodilation renders RIC ineffective in improving intestinal wall perfusion ( Fig. 5a-b) and intestinal microcirculation ( Fig. 5d-e, g-h). As a result of this, NECinduced intestinal injury and inflammation are also not improved (Fig. 6), ultimately leading to reduced survival (Fig. 7). These findings provide evidence that the RICmediated improved perfusion is required for a consequent improvement in NEC. This is addressed in the manuscript (lines 199-268).

Reviewer #3 (Remarks to the Author):
This report studies RIC in a murine model of neonatal enterocolitis. RIC is tested "early" (pre RIC days 5 and 7) and "late" (Days 6 and 8; per and post RIC) RIC is remarkably effective and the mechanism involves improved microperfusion. Improved perfusion with RIC has been shown previously in both brain and coronary circulations as the authors reference. The methods are elegant. This work is highly translatable as the authors add to the evidence that RIC is safe and effective Some questions 1.) Was RIC applied on one or both hindlimbs? Not clear in lines 305-8 We appreciate the comment from the reviewer. The RIC stimulus was always given to one hind limb and the same hind limb was used for all pups throughout the various experiments. We have clarified this in the manuscript (lines 408-409) 2.) How was this regimen of RIC chosen? They were certainly effective but how were the regimens arrived at? 9 The phenomenon of ischemic pre-conditioning was first described in the canine heart wherein four 5 min circumflex coronary occlusions, each separated by 5 min of reperfusion, dramatically reduced myocardial infarction size 29 . Przyklenk et al. also showed that brief myocardial ischemia by four cycles of 5 min coronary artery occlusion protected local and remote myocardium from sustained 1 h cardiac ischemia reperfusion injury 30 . Experimental and clinical evidence also suggests that RIC activates at least two distinct time frames of protection against ischemia reperfusion injury of the brain and heart 19,20 . The first window of protection occurs immediately after the RIC stimulus and lasts for 2 hours, and involves changes in ion channel permeability, protein phosphorylation, and release of several signaling mediators 19 . The second window of protection follows 12-24 hours after the RIC stimulus and lasts 48-72 hours, involving modulation of inflammatory response, improved endothelial function, and activation of gene expression 20 . Moreover, previous authors have reported that compared with a single episode of remote ischemic conditioning, repeated episodes were more protective in reducing inflammation in the ischemic myocardium 21 . Hence, the previously used protocol of four cycles of 5 min occlusion followed by 5 min reperfusion was chosen for RIC in our model (lines 305-319).
We selected two time periods of RIC (αRIC at P5 and P7, and βRIC at P6 and P8) which simulate two initial stages of NEC, Stage IIA and Stage IIB respectively 23,24 . The efficacy of RIC in both time periods is important for the translational application of this experimental observation. We added an explanation in the manuscript (lines 305-319).
3.) The authors studied the H2S system. The NO and NOS system has also been implicated in the mechanism of action of RIC and are important in perfusion. Were NOS 3 inhibitors tested or NOS3 KO mice? The authors might comment on this Previous authors have reported that H2S improves mesenteric perfusion and intestinal injury in experimental NEC via an eNOS-dependent mechanism 1 , suggesting that NOmediated regulation of intestinal microcirculation occurs downstream of H2S.
To further illustrate the importance of NO in RIC-mediated vasodilation during experimental NEC, we performed additional experiments using eNOS knockout mice. We demonstrated that RIC in these mice had no beneficial effect on NEC intestinal injury (morphology and inflammation) and did not improve survival. The baseline perfusion of various organs, including the intestine, was tenuous due to lack of eNOS and the measurement of intestinal wall perfusion during experimental NEC and RIC was not reliable. To overcome this difficulty, we explored the role of NO synthase inhibitor L-NAME on the effects RIC in experimental NEC. Our findings demonstrated that the RICmediated improvement in the perfusion of the whole intestinal wall and of the submucosal layer was abolished upon treatment with L-NAME. Consistent with eNOS knock out mice, inhibition of NO by L-NAME eliminated the RIC beneficial effects of reducing NEC-related intestinal injury and inflammation. The findings from these Each of my original concerns remains, and I have in re-review, identified additional concerns. Major.
1. The new concern is based upon the fact that the study does not actually provide any evidence that the intestinal microcirculation is improved in the presence of ischemic conditioning. While perfusion was enhanced (see below for the lack of proof that this is a cause and not a simple consequence of NEC improvement), the actual microcirculation i.e. the collection of blood vessels within the wall of the bowel were not shown to be altered in a way that resulted in the improvement in NEC. As such, the major premise of the work, and identified in the title even, was not supported by the data provided.
2. I take issue with their statement that the authors have shown that the restoration of intestinal perfusion is the mechanism involved. Moreover, the use of H2S scavengers -which blunt the protection of ischemic conditioning -does not logically explain an effect due to perfusion, in as much as H2S has many effects on the cell and the host, that are independent of any role on perfusion. Only by blocking perfusion specifically and losing the effects of preconditioning, can a link be made. It is true that there are effects of NO and H2S in the current studies, but these may be related to immune effects which secondarily affect perfusion, given the pleiotropic roles of these second messengers.
3. my original review included some 5 prior instances in which ischemic preconditioning reduces inflammation; I remain concerned that the work is a logical extension of the work of others. Figure 1 that the early RIC occurs actually after the first time point in the late RIC group, so the data regarding early vs. late is uninterpretable. If the authors now believe that both early and late RIC is effective, given the fact that longer durations of RIC would be expected to have greater effects on H2S or NO production, I'm even more concerned by the results. Figures 2-3 and also 5-7 are biological vs. technical vs. experimental repeats.

I remain concerned by the number of repeats and whether the data in
6. why is all RT-PCR data shown in aggregate as bars and not scatter plots? This is concerning given the high error bars.
Reviewer #2: Remarks to the Author: The authors should be commended for providing extensive new data, however some of the data does not address the initially raised concerns regarding mechanism of RIC effect, and thus overall enthusiasm for the manuscript is only modestly improved.
Top of page 15: The authors treat with H2S synthesis inhibitors and show lack of protection by RIC against NEC. They conclude that the RIC effect is due to changes in intestinal perfusion. This is not an acceptable conclusion from these data. It is highly possible that H2S is acting in some other way (than dilation) to mediate its beneficial effect since H2S has a host of other biochemical and physiological effects. Flow changes may be secondary, not causative. The problem is that the experiments necessary to prove that it changes in flow are responsible, are difficult to do. One would need to add H2S or nitric oxide and include a vasoconstrictor to prevent changes in flow to show that the beneficial effect of these agents on NEC was abrogated. In addition use of a nonspecific vasodilator such as Papaverine should be used to show that improved perfusion is sufficient to inhibit the effects of NEC. Short of such data (that dissociate changes in flow and other signaling actions of H2S and NO) the authors cannot conclude that changes in flow are mechanistically responsible. The problem with using inhibitors of H2S formation or of NO production (LNAME or eNOS KO) is that you end up blocking both the vasomotor effects and the many other biochemical signaling effects of these molecules, thus you cannot distinguish which is responsible for improvement in NEC via RPC. 1

Remote ischemic conditioning counteracts the intestinal damage of necrotizing enterocolitis by improving intestinal microcirculation
We would like to thank the editor and the reviewers for their careful review of our revised manuscript and for the thoughtful and constructive comments.
We have performed various additional experiments in both human and mice that further reinforce our initial findings. Before answering point-by-point the comments raised by the reviewers, we would like to briefly summarize (highlighted in blue) the substantial additions and changes made in the resubmitted manuscript. These stem from comments received by the reviewers as well as our motivation to make this discovery clearer and applicable to humans.
Please note that we have highlighted (in yellow) the additional experiments and modification in our resubmitted manuscript.

Summary of Added Experiments in Resubmission
1. Studies in human neonates with NEC and control neonates without NEC: These studies indicate intestinal microcirculatory deficiency in NEC.

Investigation of the role of RIC in severe NEC: This required a novel set of in vivo and in vitro experiments.
The results obtained demonstrate that RIC is beneficial in the initial stages of the disease but not when severe damage has already occurred. Clarifying the importance of the timing of RIC further supports its mechanism of action while providing important data for translation of this novel therapy into human neonates with NEC.
3. Further exploration of the effect of RIC on microcirculation by administrating nonspecific vasodilator and vasoconstrictor agents. The results obtained support our previous results obtained using eNOS knockout mice as well as chemical inhibitors and donors of NO and H 2 S. 2

Responses
Our responses to the reviewers' comments are in blue in this letter.

Reviewer #1 (Remarks to the Author):
Re: remote ischemic conditioning counteracts the intestinal damage of necrotizing enterocolitis by improving intestinal microcirculation.
I have carefully reviewed the revised manuscript.
Each of my original concerns remains, and I have in re-review, identified additional concerns. Major.
1. The new concern is based upon the fact that the study does not actually provide any evidence that the intestinal microcirculation is improved in the presence of ischemic conditioning. While perfusion was enhanced (see below for the lack of proof that this is a cause and not a simple consequence of NEC improvement), the actual microcirculation i.e. the collection of blood vessels within the wall of the bowel were not shown to be altered in a way that resulted in the improvement in NEC. As such, the major premise of the work, and identified in the title even, was not supported by the data provided.
We appreciate this concern from the reviewer. Considering the size of our experimental pups at postnatal days 5 to 9, it is not possible to isolate different areas of the intestine. Hence, the experiments necessary to directly detect blood flow and prove alterations in flow are challenging. Nonetheless, the data in our manuscript as listed below provides support that improved intestinal microcirculation due to RIC is responsible for improving the outcome of NEC: • In Figure 1, we demonstrate that human neonates with NEC have the lowest expression of vascular endothelial marker (cluster of differentiation, CD31) and highest expression of hypoxia marker (Hypoxia-inducible factor 1α, HIF1α) 1 in the most affected intestinal area. Ileum farther away from this most affected area regains similar level of CD31 expression and shows reduced HIF1α expression as compared to non-NEC control neonates. Hence, this data suggests that human NEC is associated with mucosal hypoxia and reduced number of endothelial cells suggestive of compromised intestinal perfusion.
• In Figure 2 and Figure S1b-c, we demonstrate that while RIC in wildtype mice improves the intestinal injury of NEC, reduces inflammation, and enhances survival, RIC is unable to promote the same protective effects in eNOS knockout mice. This data suggests that the protection conveyed by RIC in the intestine with NEC-induced injury is dependent on the endothelium and likely relies on endothelium-mediated vasodilation.
• Given that eNOS signaling is essential for the RIC-mediated protection against intestinal damage of NEC, we then use Doppler ultrasound to measure intestinal wall perfusion daily during the five days of our NEC induction protocol in mouse pups (Figure 3).
Intestinal wall perfusion is calculated as average flow velocity (mm/s) of multiple abdominal regions and is indicative of blood flow within the wall of the abdomen. Using these measurements, we demonstrate the following points: o As illustrated in Figure 2a and Figure S1a, there is progressive increase in intestinal injury and inflammation with significant morphological changes and elevated inflammation detectable from P7 (*p<0.05). However, as evident from Figure 3b, intestinal wall flow velocity shows significant reduction in NEC pups from P6 (*p<0.001; statistical analysis not reported in the resubmitted manuscript but can be added upon request). Hence, we detect reduced perfusion in the intestinal wall even before the intestinal epithelium is damaged. This data also demonstrates that reduced perfusion is a contributing factor to the development of NEC, rather than a consequence of NEC; this is because alterations in intestinal perfusion precedes the changes in intestinal morphology and inflammation that are associated with NEC.
o Using daily measurements of intestinal wall perfusion with Doppler ultrasound (Figure 3), we also demonstrate that in NEC pups, intestinal perfusion remains low from P5 to P9 while the non-NEC breastfed controls show increased perfusion daily. Videos S1-2 depict the derangements in intestinal wall perfusion in the NEC pups, compared to non-NEC breastfed controls. However, administration of RIC to NEC pups significantly enhances perfusion across the intestinal wall, resembling a trend similar to what is seen in the breastfed controls. Please review this data as depicted in Videos S3-4. These findings demonstrate that by manipulating intestinal blood flow with RIC during the course of NEC development, we are able to preserve perfusion and improve the outcome of NEC.
• Videos S5-8 demonstrate our evaluation of intestinal microcirculation in the submucosa using two photon laser scanning microscopy (TPLSM) which allows in vivo visualization and quantification of blood flow in real time. We have also quantified blood flow from videos obtained by TPLSM, as shown in Figure 4, to demonstrate that arteriole velocity, diameter, and flow volume in the submucosa are reduced in NEC (Please see Videos S5-6), which is suggestive of impaired perfusion. However, administration of RIC to NEC pups mitigates these derangements and preserves the velocity, diameter, and flow volume of submucosal arterioles (Please see Videos S7-8). Hence, we demonstrate that RIC improves perfusion not just across the intestinal wall as seen with Doppler ultrasound, but also in the intestinal submucosa.
• In Figure 5, we demonstrate that in agreement with preservation of submucosal perfusion by RIC, the integrity of the villi microvasculature is also improved. NEC pups demonstrated increased ischemia at the tip of the villi, evident by increased staining of the hypoxia marker pimonidazole. Consistently, NEC pups also demonstrated a marked constriction and decrease in the height of arterioles perfusing the villi. Finally, there was increased staining with the necrosis marker, Sytox Green, in the NEC pups, especially at the tip of the villi. These observations are suggestive of impairment in the integrity of intra-villi arterioles, which results in increased hypoxia at the villi tip, and hence necrosis of enterocytes. However, administration of RIC to NEC pups resulted in recovery of the diameter and height of intra-villi arterioles, resulting in reduced hypoxia and necrosis at the villi tip. This data adds to our previous findings and provides further support for the contention that RIC confers protection in the NEC intestine by mitigating the impairments in the microvasculature of the intestine.
Collectively, our data provides evidence that RIC targets the intestinal microcirculation and restores perfusion at the level of the arterioles in the villi, arterioles in the submucosa, as well as perfusion across the entire intestinal wall.
2. I take issue with their statement that the authors have shown that the restoration of intestinal perfusion is the mechanism involved. Moreover, the use of H2S scavengers -which blunt the protection of ischemic conditioning -does not logically explain an effect due to perfusion, in as much as H2S has many effects on the cell and the host, that are independent of any role on perfusion. Only by blocking perfusion specifically and losing the effects of preconditioning, can a link be made. It is true that there are effects of NO and H2S in the current studies, but these may be related to immune effects which secondarily affect perfusion, given the pleiotropic roles of these second messengers.
We appreciate the reviewer's comments regarding the mechanism of action for restoration of perfusion by RIC. We acknowledge the possibility that administration of chemical inhibitors of NO and H 2 S blocks not just the effects due to vasoregulation, but possibly other biochemical and physiological effects of these gasotransmitters as well. However, the additional experiments in our newly submitted manuscript address this concern as explained below: • As mentioned above, we tested the effectiveness of RIC in eNOS knockout pups to investigate whether the beneficial effects of RIC for avoiding the intestinal damage of NEC are dependent on the endothelium. In Figure 2 and Figure S1b-c, we demonstrate that unlike in wildtype mice, administration of RIC to eNOS knockout pups with NEC fails to improve intestinal morphology, reduce inflammation, or enhance survival. This data provides evidence that the endothelium plays an essential role in the mechanism of action of RIC.
• In Figure S7, we demonstrate that following administration of methoxamine, a general and nonspecific vasoconstrictor 2 , the protective effects of RIC in the NEC intestine are lost. This additional data provides further support for the beneficial effects of RIC due to regulation of blood flow via vasodilation.
• In Figure S7, we also demonstrate that administration of two nonspecific vasodilators, papaverine and captopril, to NEC pups improves intestinal injury, reduces inflammation, and enhances survival; hence providing the same protective effects conferred by RIC. These findings provide evidence that vasodilation is sufficient to improve the outcome of NEC.
3. my original review included some 5 prior instances in which ischemic preconditioning reduces inflammation; I remain concerned that the work is a logical extension of the work of others.
We appreciate the reviewer's concern regarding the effects of ischemic conditioning on inflammation. Data in our newly submitted manuscript provides evidence that changes in intestinal blood flow occur prior to inflammation and injury in the epithelium (Figure in this  letter). Hence, derangements of blood flow play a primary role in the development of NEC. Our evidence for this contention is listed below: • We have previously demonstrated that in the early neonatal period (P5), pups fail to stimulate an increase in intestinal blood flow in response to a single gavage formula feeding 3 . This inadequate response to feeding stems from the immaturity of the intestinal microvasculature and prevents the intestine from meeting its increased oxygen demand after feeding 3 . Hence, feeding continues to disturb the balance between oxygen demand and supply in the premature intestine, leading to hypoxia, and the development of NEC 3 . These previous findings contribute to our understanding of the emergence of the impairments in intestinal microcirculation and hypoxia that are known to be associated with NEC 4-8 . Collectively, the alterations in intestinal blood flow dynamics which arise due to prematurity and feeding-induced hypoxia are primary events in the development of NEC. Moreover, alterations in intestinal blood flow occur prior to any changes in intestinal morphology or inflammation.
In the current study, in Figure S2a-c, using TPLSM, we demonstrate that administration of RIC to P5 pups offsets the poor microcirculatory response to feeding. Administration of RIC to pups in the early neonatal period enables increased blood flow in the intestinal 6 microvasculature following feeding, resembling the response seen in mature pups at P9. Quantification of blood flow from videos obtained in real time by TPLSM revealed enhanced velocity, diameter, and flow volume of submucosal arterioles following feeding of P5 pups that received RIC. Hence, we show that by modulating the immature intestinal microcirculation and improving blood flow at an early stage, we can prevent the development of NEC. This data provides further evidence that modulating the blood flow dynamics in the intestine, even before any injury or inflammation has occurred, plays a primary role in improving the outcome of the disease.
• In our resubmitted manuscript, we added various experiments to evaluate whether RIC is still effective if administered in late stages of NEC development, when significant injury and inflammation in the intestine have already been established. As mentioned above in our response to the first comment, morphological changes and increase in inflammation are first detected at P7 during the development of NEC. In Figure 2 of the resubmitted manuscript and the Figure in this letter, we demonstrate that administration of Stage 3 RIC at this point (P7), when NEC-induced injury and inflammation in the intestine are already established, is not able to reverse the injury and inflammation or convey protection. For this reason, Stage 3 RIC also failed to enhance survival of pups, unlike Stage 1 or 2 RIC which are administered at early stages of the disease to confer protection. Hence, it is unlikely that inflammation is the primary target of RIC for protecting the intestine.
• In the newly added experiments in our manuscript, we have also investigated the effects of nonspecific vasodilators, papaverine and captopril, in NEC pups. These drugs directly affect the regulation of blood flow, with no reports suggesting an effect on inflammation 9 . By demonstrating that administration of these vasodilators alone is sufficient to convey the same protective effects of RIC and improve the outcome of NEC, we provide further support that reduced inflammation is only a secondary outcome of RIC.
We appreciate the evidence provided by previous studies in the literature that RIC reduces inflammation. However, we hope that with our additional experiments, we are able to convince the reviewer that in this case, mitigation of inflammation is secondary to the preservation of intestinal blood flow in the mechanism of action of RIC for improving NEC outcome.
4. I remain very concerned by the data in Figure 1 that the early RIC occurs actually after the first time point in the late RIC group, so the data regarding early vs. late is uninterpretable. If the authors now believe that both early and late RIC is effective, given the fact that longer durations of RIC would be expected to have greater effects on H2S or NO production, I'm even more concerned by the results.
We appreciate the reviewer's concern regarding the selected time points for the administration of RIC. Please allow us to address this confusion by illustrating a few points: 7 • In our revised manuscript, we updated the terminology of early and late RIC to Stage 1 and Stage 2 RIC, respectively. Please allow us to adhere to this terminology as it avoids further confusion.
• • Previous reports suggest that in addition to short-lasting protective effects that are conferred by RIC immediately after it is administered, RIC also activates a time window of protection which occurs 12-24 hours after it is administered, which lasts for 48-72 hours 10,11 . This is why for each Stage of RIC (1, 2, 3) we administered RIC in two episodes and on two non-consecutive days during the 5-day period of NEC induction; to ensure that the second window of protection activated by RIC as suggested by the literature remains in effect until the day of sacrifice. By this logic: o Stage 1 RIC involves a first episode of RIC given on the first day of NEC induction (P5), followed by a second episode of RIC given 48 hours later, on the third day of NEC induction (P7).
o Stage 2 RIC involves a first episode of RIC given on the second day of NEC induction (P6), followed by a second episode of RIC given 48 hours later, on the fourth day of NEC induction (P8).
o Stage 3 RIC involves a first episode of RIC given on the third day of NEC induction (P7), followed by a second episode of RIC given 48 hours later, on the final day of NEC induction (P9).
• The importance of the chosen timepoints for RIC as explained above underlies in its implication in the clinical setting. The prospect that we have for application of RIC in the clinical setting is to implement it for treatment of three groups of neonates with NEC: 8 o Neonates with Stage I NEC which present with suspected or minimal intestinal damage [12][13][14] o Neonates with Stage II NEC, which present with moderate intestinal damage [12][13][14] o Neonates with Stage III NEC, which present with severe intestinal damage [12][13][14] In all of these patient groups, one single episode of RIC will not be effective for preventing the progressing of the disease and conferring protection. This stems from the explanation given above regarding the different time windows of protection activated by RIC. Hence, in the clinical setting, RIC will have to be administered on multiple days. • Regarding experimental repeats, all experiments include mice taken from various litters. 6. why is all RT-PCR data shown in aggregate as bars and not scatter plots? This is concerning given the high error bars.
We can modify the data in the manuscript as requested to display our qRT-PCR data using scatter plots.

Reviewer #2 (Remarks to the Author):
The authors should be commended for providing extensive new data, however some of the data does not address the initially raised concerns regarding mechanism of RIC effect, and thus overall enthusiasm for the manuscript is only modestly improved.
Top of page 15: The authors treat with H2S synthesis inhibitors and show lack of protection by RIC against NEC. They conclude that the RIC effect is due to changes in intestinal perfusion. This is not an acceptable conclusion from these data. It is highly possible that H2S is acting in some other way (than dilation) to mediate its beneficial effect since H2S has a host of other biochemical and physiological effects. Flow changes may be secondary, not causative. The problem is that the experiments necessary to prove that it changes in flow are responsible, are difficult to do. One would need to add H2S or nitric oxide and include a vasoconstrictor to prevent changes in flow to show that the beneficial effect of these agents on NEC was abrogated.
In addition use of a nonspecific vasodilator such as Papaverine should be used to show that improved perfusion is sufficient to inhibit the effects of NEC. Short of such data (that dissociate changes in flow and other signaling actions of H2S and NO) the authors cannot conclude that changes in flow are mechanistically responsible. The problem with using inhibitors of H2S formation or of NO production (LNAME or eNOS KO) is that you end up blocking both the vasomotor effects and the many other biochemical signaling effects of these molecules, thus you cannot distinguish which is responsible for improvement in NEC via RPC.
We appreciate the comments from the reviewer. The constructive suggestions allowed us to conduct additional experiments to improve our manuscript and provide further support for our hypothesis.
• To prove that changes in intestinal blood flow are sufficient to improve the outcome of NEC, we followed the reviewer's suggestions to investigate the effects of nonspecific vasodilators. In Figure S7a-c, we demonstrate that administration of papaverine and captopril improves intestinal injury, reduces inflammation, and enhances survival. Using these nonspecific vasodilators, we have dissociated the changes in flow from other biochemical and physiological effects of NO and H 2 S. Our current data demonstrates that preservation of intestinal microcirculation is responsible for conferring the RIC-mediated protection and reversing the intestinal damage of NEC.
• Additionally, we have investigated the effects of methoxamine, an intestinal vasoconstrictor 2 in the presence of RIC. As shown in Figure S7d, methoxamine abolished the beneficial effects of RIC in prolonging survival, which is the most crucial aspect of RIC-mediated protection in NEC from a clinical perspective. This data provides further support that vasodilation is essential to the ability of RIC to mitigate the intestinal damage of NEC and improve disease outcome.
We hope that with these additional experiments, we are able to convince the reviewer of the mechanism of action of RIC being targeted primarily towards the intestinal microcirculation.

Reviewer #3 (Remarks to the Author):
This report studies RIC in a murine model of neonatal enterocolitis. RIC is tested "early" (pre RIC days 5 and 7) and "late" (Days 6 and 8; per and post RIC) RIC is remarkably effective and the mechanism involves improved microperfusion. Improved perfusion with RIC has been shown previously in both brain and coronary circulations as the authors reference. The methods are elegant. This work is highly translatable as the authors add to the evidence that RIC is safe and effective Some questions 1.) Was RIC applied on one or both hindlimbs? Not clear in lines 305-8 We appreciate the comment from the reviewer. The RIC stimulus was always given to one hind limb and the same hind limb was used for all pups throughout the various experiments. We have clarified this in the manuscript.
2.) How was this regimen of RIC chosen? They were certainly effective but how were the regimens arrived at?
The phenomenon of ischemic pre-conditioning was first described in the canine heart wherein four 5 min circumflex coronary occlusions, each separated by 5 min of reperfusion, dramatically reduced myocardial infarction size 16 . Przyklenk et al. also showed that brief myocardial ischemia by four cycles of 5 min coronary artery occlusion protected local and remote myocardium from sustained 1 h cardiac ischemia reperfusion injury 17 . Experimental and clinical evidence also suggests that RIC activates at least two distinct time frames of protection against ischemia reperfusion injury of the brain and heart 10,11 . The first window of protection occurs immediately after the RIC stimulus and lasts for 2 hours, and involves changes in ion channel permeability, protein phosphorylation, and release of several signaling mediators 10 . The second window of protection follows 12-24 hours after the RIC stimulus and lasts 48-72 hours, involving modulation of inflammatory response, improved endothelial function, and activation of gene expression 11 . Moreover, previous authors have reported that compared with a single episode of remote ischemic conditioning, repeated episodes were more protective in reducing inflammation in the ischemic myocardium 15 . Hence, the previously used protocol of four cycles of 5 min occlusion followed by 5 min reperfusion was chosen for RIC in our model. We selected two time periods of RIC (Stage 1 RIC at P5 and P7, and Stage 2 RIC at P6 and P8) which simulate two initial stages of NEC, Stage IIA and Stage IIB respectively 13,14 . The efficacy of RIC in both time periods is important for the translational application of this experimental observation. We added an explanation in the manuscript.
3.) The authors studied the H2S system. The NO and NOS system has also been implicated in the mechanism of action of RIC and are important in perfusion. Were NOS 3 inhibitors tested or NOS3 KO mice? The authors might comment on this Previous authors have reported that H2S improves mesenteric perfusion and intestinal injury in experimental NEC via an eNOS-dependent mechanism1, suggesting that NO-mediated regulation of intestinal microcirculation occurs downstream of H2S 18 . To further illustrate the importance of NO in RIC-mediated vasodilation during experimental NEC, we performed additional experiments using eNOS knockout mice. We demonstrated that RIC in these mice had no beneficial effect on NEC intestinal injury (morphology and inflammation) and did not improve survival. The baseline perfusion of various organs, including the intestine, was tenuous due to lack of eNOS and the measurement of intestinal wall perfusion during experimental NEC and RIC was not reliable (Supplementary Figure S3). To overcome this difficulty, we explored the role of NO synthase inhibitor L-NAME on the effects RIC in experimental NEC. Our findings demonstrated that the RIC-mediated improvement in the perfusion of the whole intestinal wall and of the submucosal layer was abolished upon treatment with L-NAME. Consistent with eNOS knock out mice, inhibition of NO by L-NAME eliminated the RIC beneficial effects of reducing NEC-related intestinal injury and inflammation. The findings from these additional experiments can be found in the results, and in Fig. 2g-h, Fig. 6, Fig. 7, and supplementary Fig. S1c, and Fig. S6.

Final considerations
We would like to thank the editors for giving us the chance to clarify the comments and concerns of the reviewers. This is a valuable opportunity and we hope that we have been able to provide further clarification for the remaining concerns of the reviewers.
The main concern of Reviewer #1 seemed to be related to whether the changes in intestinal perfusion conferred by RIC precedes changes in intestinal inflammation. We hope that our explanations can convince the reviewer that while RIC has been shown to primarily target inflammatory cascades in previous studies, in our study, the primary effect of RIC is preservation of intestinal microcirculation, followed by decreased injury and inflammation in the intestine, enhanced survival, and hence improvement of NEC outcome.
We were happy that after our modifications and added experiments in the revised manuscript, Reviewer #2 agreed that RIC confers its protective effects in the NEC intestine by improving perfusion across the entire intestinal wall, as well as in the submucosal arterioles and within the villi. In our resubmitted manuscript, we followed the advice of the reviewer and performed additional experiments to further investigate the role of vasodilation in the RIC-mediated effects on intestinal blood flow dynamics. We hope that by investigating the effects of nonspecific vasodilator and vasoconstrictor agents, we have been able to show that vasodilation alone is sufficient to confer the protection of RIC in the NEC intestine.
Finally, we assume that we have been successful in addressing all of the comments and concerns of Reviewer #3, who had reviewer our original manuscript, as no further comments were provided by this reviewer.
Thank you very much for your consideration of our manuscript for publication in Nature Communications.